Making decision for investment is often concerned with prediction of data’s trend or value. To predict trend or value, we must use some techniques for analyzing data to discover patterns. Then we use that discovered patterns for prediction. Data mining using backpropagation neural network is one solution to prediction. It’s very interesting and can be adapted in many application. This project present the creation of Backpropagation neural network generator and trainer program using object-oriented programming, and experimentation aimed to find the best topology for neural network which will be used to predict the lottery.